Bank restricts AI agent access to payment functions

The measure targets AI agents operated outside the bank, preventing them from executing money transfers and payments through JPMorgan's systems. The move reflects growing concern among financial institutions that autonomous AI agents, which can act on their own without direct human involvement, could be exploited for unauthorized transactions, fraud, or other abuses if granted access to payment infrastructure.

Broader context: banks guard against autonomous AI risk

JPMorgan's decision comes as agentic AI — systems designed to plan and execute tasks independently — spreads rapidly across consumer and enterprise software. Payment and remittance functions represent a particularly sensitive area, since agent-driven transactions carry risks of erroneous transfers and account compromise that differ from traditional automated banking services.

By blocking external agents at the transaction layer, the bank is effectively requiring that payment activity be tied to human-directed, authenticated sessions rather than to third-party AI software acting on a user's behalf. The approach signals that major banks are willing to restrict emerging agent capabilities outright rather than wait for security frameworks and regulations to mature.

The judgment on this claim is: True